Energy consumption and surface roughness modelling for multi-objective optimisation of machining processes

Sustainability and precision are fundamental requirements for machining processes. In this study, novel energy consumption and surface roughness models are established to support the multi-objective (sustainability and precision) optimisation of machining processes. In this study, the following innovative characteristics are exhibited: (i) a physically interpretable energy consumption model is designed by considering correlations between key machining parameters and machining powers used in different machining zones; (ii) a surface roughness model is designed based on combined machine learning models, and optimised weights for the combination are achieved using the genetic algorithm; (iii) based on the energy consumption and surface roughness models, a Pareto-based multi-objective optimisation algorithm is devised to obtain optimised machining parameters. Orthogonal experiments were conducted to validate the models and algorithm. Results show that the energy consumption model reached an average absolute percentage error of only 1.7% and an R² determination coefficient of 99.1%. In comparison with other models, the surface roughness model designed in this study achieved the best performance (the mean absolute percentage error (MAPE) reached 2.5%, and the R² coefficient was above 95%). The Pareto optimal solution set for machining parameters was obtained efficiently.